Search

Extracting planting information of rice as early as possible can be great significance in provincial agricultural production for the food security. MODIS data has been demonstrated to be superior in extracting planting information of rice at large scale due to short observation period and wide swath and easy image acquisition. The method of MODIS index during early growth period of rice combined with threshold value was usually used in the conventional provincial decision-making service to ensure the timeliness of service and convenient operation. However, the method is largely influenced by human and phenology difference of rice in various regions leading to highly subjective and poor stability. In addition, the mixed pixels were likely causing misestimation of the MODIS product. Random forest algorithm can make up the deficiency of threshold method due to the characteristics of less manual intervention and difficult over-fitting. In the study, an extraction method of rice planting information was proposed using MODIS index and random forest during early growth period of rice. Firstly, Enhanced Vegetation Index (EVI), Normalized Difference Vegetation Index (NDVI) and Land Surface Water Index (LSWI) were selected to extract planting information based on the variation characteristics of MODIS index from seeding to jointing stage in Jiangxi Province, and then feature data sets were constructed to model. The models were built to extract planting area of early rice and to inverse planting abundance using random forest algorithm. Finally, the accuracy of the planting and the abundance maps of early rice were validated by the verification samples from the measured points, validation sample region obtained by the Sentinel-1A image and the statistical data from Jiangxi Provincial Bureau of Statistics. The results showed that using MODIS index during early growth period of rice and random forest was an effective way to extract plant information. The classification accuracy of early rice planting area extraction model was 93.18% with the Kappa coefficient of 0.915, and the mapping accuracy and user accuracy were 92.04% and 91.23%, respectively when matching the verification samples. Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), decision coefficient of abundance inversion model were 0.07, 0.104 and 0.855, respectively, and the better performance of the abundance inversion model was achieved in high abundance planting areas. The spatial distribution characteristics of early rice planting area and abundance were consistent with validation sample region. Compared with the statistical data, the model accuracy of early rice area was 92.33%. This method can ensure the timeliness of the service and solve the problem that the extraction of rice planting area in the conventional provincial decision-making service is greatly affected by the problems of rationality of threshold selection and mixed pixel, also have no complex operation. The finding can provide a reference to extract planting area of early rice during early growth period. [ABSTRACT FROM AUTHOR]

With the continuous development of hyperspectral remote sensing technology, it has been widely used in vegetation mapping. However, sparse vegetation canopy, soil background, and spectral similarity between different types of vegetation are still the main challenges for vegetation types mapping in arid areas. As a result, it is difficult to achieve reliable classification accuracy by using spectral or texture features separately. Generalized Normal Distribution Optimization (GNDO) is a new feature optimization algorithm, with advantages in quality and stability of feature extraction results, comparing to traditional optimization algorithms. But it has not yet been applied to select bands of hyperspectral data. In order to validate the feasibility of combining ZY-1 02D spectral and texture features to classify vegetation types in arid areas, to verify the effectiveness of the GNDO method for bands selection of hyperspectral data, and to explore the effects of feature selection methods and training sample numbers on the classification accuracy of vegetation mapping, different Wrapper Optimization methods, such as Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Grey Wolf Optimization (GWO), and GNDO, were applied to select spectral features for vegetation mapping, taking the area around Zongjia Town, Dulan County, Qinghai Province, China as the research area, then the band selection results of these methods were analyzed. Train sample set containing 30, 50, 100, 150, and 200 pixels per class were used to select bands and to train the classifier. Different methods (ALL (without bands selection), GA, PSO, GWO, GNDO) and different sizes of the trained sample sets (30, 50, 100, 150, and 200 pixels per class) were used to obtain 25 spectral feature-based classification data sets. Simultaneously, 8 texture features (mean, variance, homogeneity, contrast, dissimilarity, entropy, second moment) were extracted using the Gray-level co-occurrence matrix (GLCM) method, and selected on basis of distinguishability for vegetation types. Texture features (TEX) were combined with spectral feature-based classification data sets. The random forest classification method was applied to classify vegetation types for the classification data sets, and the classification accuracy of classification data sets was evaluated and compared. The results show that 1) the blue region (400-450 nm), the red edge region (700-750 nm), and the red region (600-650 nm) are the most sensitive to distinguish the vegetation types in the study area; 2) the GNDO200 achieved the highest overall classification accuracy (80.44%) among the spectral feature-based classification data sets, which was better than the classification accuracy (78.86%) using all bands (ALL200); 3) with the increase of training samples, the overall classification accuracy of each classification data set showed an increasing trend, the classification accuracy of different feature selection methods showed different reliance on the number of training samples; 4) image texture features significantly improved the classification accuracy, and the GWO200+TEX dataset had the highest overall classification accuracy (82.86%). This study could verify the potential of the ZY1-02D, the new hyperspectral satellite data, for the classification of vegetation types in arid areas, and provide an idea for the selection of spectral and texture features in hyperspectral vegetation mapping. [ABSTRACT FROM AUTHOR]